Serum autotaxin is increased in pruritus of cholestasis, but not of other origin, and responds to therapeutic interventions


  • Andreas E. Kremer,

    1. Tytgat Institute for Liver and Intestinal Research, Department of gastroenterology and Hepatology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands
    2. Department of Medicine I, Friedrich-Alexander-University of Erlangen-Nuremberg, Erlangen, Germany
    Search for more papers by this author
  • Remco van Dijk,

    1. Tytgat Institute for Liver and Intestinal Research, Department of gastroenterology and Hepatology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands
    Search for more papers by this author
  • Pamela Leckie,

    1. Upper Third, UCL Medical School, Royal Free Hospital, London, United Kingdom
    Search for more papers by this author
  • Frank G. Schaap,

    1. Tytgat Institute for Liver and Intestinal Research, Department of gastroenterology and Hepatology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands
    Search for more papers by this author
  • Edith M.M. Kuiper,

    1. Department of Gastroenterology and Hepatology, Erasmus MC University Medical Center, Rotterdam, The Netherlands
    Search for more papers by this author
  • Thomas Mettang,

    1. Department of Nephrology, German Clinic for Diagnostics, Wiesbaden, Germany
    Search for more papers by this author
  • Katrin S. Reiners,

    1. Laboratory of Immunotherapy, Department I of Internal Medicine, University Clinic Cologne, Cologne, Germany
    Search for more papers by this author
  • Ulrike Raap,

    1. Department of Dermatology and Allergy, Hannover Medical School, Hannover, Germany
    Search for more papers by this author
  • Henk R. van Buuren,

    1. Department of Gastroenterology and Hepatology, Erasmus MC University Medical Center, Rotterdam, The Netherlands
    Search for more papers by this author
  • Karel J. van Erpecum,

    1. Department of Gastroenterology and Hepatology, University Medical Center, Utrecht, The Netherlands
    Search for more papers by this author
  • Nathan A. Davies,

    1. Upper Third, UCL Medical School, Royal Free Hospital, London, United Kingdom
    Search for more papers by this author
  • Christian Rust,

    1. Department of Medicine II, Klinikum Grosshadern, University of Munich, Munich, Germany
    Search for more papers by this author
  • Andreas Engert,

    1. Chairman, German Hodgkin Study Group, Department I of Internal Medicine, University Hospital of Cologne, Cologne, Germany
    Search for more papers by this author
  • Rajiv Jalan,

    1. Upper Third, UCL Medical School, Royal Free Hospital, London, United Kingdom
    Search for more papers by this author
  • Ronald P.J. Oude Elferink,

    1. Tytgat Institute for Liver and Intestinal Research, Department of gastroenterology and Hepatology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands
    Search for more papers by this author
  • Ulrich Beuers

    Corresponding author
    1. Tytgat Institute for Liver and Intestinal Research, Department of gastroenterology and Hepatology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands
    • Department of Gastroenterology and Hepatology, Tytgat Institute for Liver and Intestinal Research, Academic Medical Center, G4-216, University of Amsterdam, P.O. Box 22600, NL-1100 DD Amsterdam, The Netherlands
    Search for more papers by this author
    • fax: +31-20-566 95 82

  • Potential conflict of interest: A.E.K., R.O.E., and U.B. have filed a patent on autotaxin (WO 2011/049433 A1).

  • The authors thank Dr. J. Aoki (University of Tokyo, Tokyo, Japan) for kindly providing 4F1/autotaxin antibody and Dr. R. Hoekstra (University of Amsterdam, Amsterdam, The Netherlands) for generously providing HepG2 cells overexpressing pregnane X receptor.


Pruritus is a seriously disabling symptom accompanying many cholestatic liver disorders. Recent experimental evidence implicated the lysophospholipase, autotaxin (ATX), and its product, lysophosphatidic acid (LPA), as potential mediators of cholestatic pruritus. In this study, we highlight that increased serum ATX levels are specific for pruritus of cholestasis, but not pruritus of uremia, Hodgkin's disease, or atopic dermatitis. Treatment of patients with cholestasis with the bile salt sequestrant, colesevelam, but not placebo, effectively reduced total serum bile salts and fibroblast growth factor 19 levels, but only marginally altered pruritus intensity and ATX activity. Rifampicin (RMP) significantly reduced itch intensity and ATX activity in patients with pruritus not responding to bile salt sequestrants. In vitro, RMP inhibited ATX expression in human HepG2 hepatoma cells and hepatoma cells overexpressing the pregnane X receptor (PXR), but not in hepatoma cells in which PXR was knocked down. Treatment of severe, refractory pruritus by the molecular adsorbents recirculation system or nasobiliary drainage improved itch intensity, which, again, correlated with the reduction of ATX levels. Upon reoccurrence of pruritus, ATX activity returned to pretreatment values. Conclusion: Serum ATX activity is specifically increased in patients with cholestatic, but not other forms of, systemic pruritus and closely correlates with the effectiveness of therapeutic interventions. The beneficial antipruritic action of RMP may be explained, at least partly, by the PXR-dependent transcriptional inhibition of ATX expression. Thus, ATX likely represents a novel therapeutic target for pruritus of cholestasis. (HEPATOLOGY 2012)

(See Editorial on Page 1194)

Chronic pruritus can be a seriously debilitating symptom accompanying various cutaneous and systemic disorders.1 It represents one of the most prominent clinical features in numerous liver disorders, such as primary biliary cirrhosis (PBC), primary sclerosing cholangitis (PSC), cholangiocarcinoma (CCC), inherited forms of cholestasis, and intrahepatic cholestasis of pregnancy.2 This form of itching is designated cholestatic pruritus because impaired bile flow is a common denominator in these disorders.3 The molecular mechanisms involved in the pathogenesis of cholestatic pruritus remain enigmatic, and treatment of these patients often represents a clinical challenge because of limited therapeutic options. The current guidelines for the treatment of cholestatic pruritus recommend the use of bile salt sequestrants, such as cholestyramine or colesevelam, as first-line therapy and rifampicin (RMP) as second-line treatment.2, 4 If patients do not respond to these drugs, experimental approaches may be applied. The molecular absorbance recirculating system (MARS) is an extracorporeal liver dialysis system capable of removing mainly albumin-bound molecules, such as bile salts, bilirubin, ammonia, and other amphiphilic toxins. MARS therapy has been shown to effectively alleviate intractable pruritus of cholestasis in patients who do not respond to any medicinal therapy.2, 4, 5 Nasobiliary drainage transiently relieves severe pruritus in benign recurrent intrahepatic cholestasis (BRIC)6 and primary biliary cirrhosis (PBC) patients7 who do not respond to standard antipruritic treatment.2, 4 However, pruritus may even become refractory to all medical treatments and can be an indication for liver transplantation, even in the absence of liver failure.4

By functional screening of sera of patients with cholestasis suffering from pruritus on neuronal cells, we recently identified lysophosphatidic acid (LPA) as a potent neuronal activator.8 Serum levels of this phospholipid were increased in patients with cholestasis that suffered from pruritus. Circulating LPA is formed by a lysophospholipase D, called autotaxin (ATX), which hydrolyzes the choline group from lysophosphatidylcholine (LPC).9 In mice, the amount of circulating LPA depends on serum ATX activity.10 In line with the observed increase in LPA, ATX activity was higher in sera of patients with pruritus with cholestatic disorders, compared to those without pruritus. Furthermore, itch intensity highly correlated with ATX activity. Intradermal injection of LPA caused a dose-dependent scratch response in mice.8

ATX was initially identified as a cell motility factor, which is overexpressed in various tumors and is involved in the proliferation and generation of metastases.11 The effects of ATX are largely mediated by the enzymatic formation of LPA, which activates at least six different G-protein-coupled receptors.9, 11 ATX is also essential for angiogenesis, neuronal development, and lymphocyte homing,10 and LPA mediates the initiation of neuropathic pain, hair growth, and embryo implantation.9

Here, we studied whether increased serum ATX activity is specific for pruritus of cholestasis. We also aimed to investigate the effect of various therapeutic interventions, such as treatment with colesevelam, rifampicin (RMP), MARS, and nasobiliary drainage on ATX activity. Finally, the effects of RMP on ATX expression were studied in vitro.


ANOVA, analysis of variance; ATX, autotaxin; BRIC, benign recurrent intrahepatic cholestasis; BSA, bovine serum albumin; CCC, cholangiocarcinoma; ChO, choline oxidase; DMEM, Dulbecco's modified Eagle's medium; DMSO, dimethyl sulfoxide; FGF, fibroblast growth factor; HL, Hodgkin's lymphoma; HRP, horseradish peroxidase; HVA, homovanillinic acid; ISS, itch severity score; LPA, lysophosphatidic acid; LPC, lysophosphatdiylcholine; MARS, molecular absorbance recirculating system; PBC, primary biliary cirrhosis; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; PPV, positive predictive value; PSC, primary sclerosing cholangitis; PUO, pruritus of unknown origin; PXR, pregnane X receptor; qPCR, quantitative polymerase chain reaction; RMP, rifampicin; ROC, receiver operator characteristic; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis; shRNA, short hairpin RNA; TBS, total serum bile salts; VAS, visual analog scale.

Materials and Methods

Human Subjects.

Peripheral venous blood was obtained from healthy donors and patients with cholestatic disorders, uremia, Hodgkin's disease, and atopic dermatitis after informed consent, according to the Declaration of Helsinki. The study was approved by the local medical ethical committees. Treatment interventions, such as colesevelam,12 RMP, MARS therapy, and nasobiliary drainage,7 were conducted, recorded, and reported on in compliance with the International Conference on Harmonization Good Clinical Practice and national regulations. Blood samples were allowed to clot for 1 hour before they were centrifuged at 4°C, and serum was cryopreserved in aliquots at −80°C. Itch intensity was quantified in all patients at the time point of blood drawing using a visual analog scale (VAS) ranging from 0 (no pruritus) to 100 (unbearable pruritus). In the colesevelam study,12 35 patients were evaluable, of whom 17 patients received colesevelam (1,875 mg twice-daily) and 18 patients were treated with an identical placebo for 3 weeks. The study population consisted of 22 female and 13 male patients being mainly diagnosed for PBC (N = 14) or PSC (N = 14). MARS treatment was performed in 10 patients (8 female and 2 male) with intractable pruritus resulting from PBC (n = 6), PSC (n = 2), or other liver disorders (n = 2; Supporting Table 4).


Choline oxidase (ChO), horseradish peroxidase (HRP), homovanillinic acid (HVA), dimethyl sulfoxide (DMSO), bovine serum albumine (BSA), and RMP were purchased from Sigma-Aldrich (Steinheim, Germany); stearoyl LPA (18:1) and myristoyl LPC (14:0) were from Avanti Lipids (Alabaster, AL).

Cell Culture.

Human HepG2 hepatoma cells were grown in Dulbecco's modified Eagle's medium (DMEM; Lonza BioWhittaker, Cologne, Germany) supplemented with 10% fetal calf serum, 4 mM of L-glutamine, and a mixture of antibiotics (5 mg/mL of penicillin and 5 mg/mL of streptomycin). Cells were incubated at 37°C in a humidified atmosphere containing 5% CO2. For studying the effect of RMP, cells were seeded in six-well plates at a density of 8 × 105 cells/well until reaching 80% confluence. Subconfluent cells were cultured overnight in serum-free medium containing 0.2% BSA. After brief washing, cells were incubated for 24 hours in DMEM/0.2% BSA containing 10 μM of RMP. As a solvent control, 0.1% DMSO was added to control cells. HepG2 cells overexpressing PXR and PXR knock-down HepG2 cells (see below) were identically analyzed.

Lentiviral Transduction.

Short hairpin RNAs (shRNAs) for PXR (TRCN0000021623) and plasmids encoding nontarget control (SHC002) as the control were obtained from Sigma-Aldrich in a lentiviral vector system. Lentiviral supernatant was generated from HEK 293T cells as packaging cells (calcium-phosphate–based transfection; ClonTech, Mountain View, CA). HepG2 cells were grown to 50%-60% confluence and incubated with virus-containing supernatants/DMEM (1:1) supplemented with 10 μg/mL of diethylaminoethyl/dextrane for 24 hours. Selection of transduced cells was achieved by the addition of puromycin (30 μg/mL), and knockdown of PXR was verified by quantitative polymerase chain reaction (qPCR) (see below). HepG2 cells overexpressing PXR were generously provided by Dr. R. Hoekstra (Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands).13

RNA Isolation and Quantification of Transcript Levels.

Total RNA was extracted from cultured cells using Trizol reagent (Invitrogen, Carlsbad, CA). Complementary DNA was synthesized from total RNA with an oligo-dT primer and Superscript III reverse transcriptase (Invitrogen). Real-time PCR measurements were performed at 60°C in a Lightcycler apparatus (Roche, Mannheim, Germany) with Lightcycler Faststart DNA Master Plus CYBR Green I (Roche). Transcript levels were normalized to the housekeeping gene, 36b4 (acidic ribosomal phosphoprotein P0). For qPCR experiments, the following primer sequences were used: ATX forward: TGCAATAGCTCAGAGGACGA; ATX reverse: AGAAGTCCAGGCTGGTGAGA; CYP3A4 forward: TGTTTTCAGCCCATCTCCTT; CYP3A4 reverse: CATTGCATCGAGACAGTTGG; 36B4 forward: TCATCAACGGTACAAACGA; and 36B4 reverse: GCCTTGACCTTTTCAGCAAG.

ATX Activity Assay.

ATX activity was quantified in diluted sera, as recently described.8 Briefly, serum samples were incubated with a buffer containing 500 mmol/L of NaCl, 5 mmol/L of MgCl2, 100 mmol/L of Tris (pH = 9.0), and 0.05% Triton X-100 for 60 minutes at 37°C. Parallel incubations were performed in the presence and absence of 1 mmol/L of LPC (14:0). The lysophospholipase activity of ATX was determined by the amount of liberated choline, as detected by enzymatic fluorimetry using ChO (2 U/mL), HRP (1.6 U/mL), and HVA as substrates for peroxidase. After the addition of both enzymes in a buffer (consisting of 20 mmol/L of CaCl2, 2 mmol/L of HVA, 50 mmol/L of 3-[N-morpholino]propanesulfonic acid [pH = 8.0], and 0.1% Triton X-100), the increase in fluorescence was monitored at 37°C on a NOVOstar analyzer (excitation 320 nm and emission 405 nm; BMG Labtech GmbH, Offenburg, Germany). The (endogenous) amount of choline present in the sample without the addition of LPC was subtracted from the amount measured in the presence of LPC. Interassay variance of the assay was less than 15%, and intraassay variance of the assay was below 10%. For studying the effect of RMP on in vitro ATX activity, healthy control serum was incubated using the above-mentioned buffers containing different concentrations of RMP. Stock solutions of RMP were dissolved in phosphate-buffered saline (PBS) and were diluted 100-fold in the above-mentioned buffer. PBS was used as a solvent control.

Determination of Fibroblast Growth Factor 19.

Serum fibroblast growth factor (FGF)-19 levels were determined using a sandwich enzyme-linked immunosorbent assay specific for FGF-19, as described recently.14

Determination of Bile Salts.

Total serum bile salt levels were quantified using a Diazyme total bile salts kit (Diazyme Laboratories, Poway, CA), according to the manufacturer's instructions.

Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis and Western Blotting.

Serum samples or albumin dialysates were diluted and incubated for 10 minutes at 37°C with sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) loading buffer containing β-mercapto-ethanol. Amounts corrected for protein content were separated by SDS-PAGE, blotted on polyvinylidene fluoride membranes, blocked with 5% skim milk in PBS, and incubated with a rat antihuman ATX antibody (monoclonal antibody 4F1, 1:10,000; kindly provided by J. Aoki)15 and appropriate secondary detection reagents. Immunoreactive bands were visualized by enhanced chemoluminescence (Roche, Amersham, Buckinghamshire, UK).

Statistical Analysis.

Statistical differences were evaluated for two groups by the Student t test and for three or more groups by one-way analysis of variance (ANOVA) with Bonferroni correction using SPSS (version 18.0; SPSS, Inc., Chicago, IL). A paired t test was used if values before and after therapy were compared. Pearson's correlation coefficient and corresponding P values were calculated to assess the relationship between tested parameters. A multivariable test score was constructed from a logistic regression model, with disease status as the dependent variable and ATX as the independent variable. Test performance was then assessed by calculating the c-statistic (i.e., area under the receiver operating characteristic [ROC]). All data are expressed as means ± standard deviations.


Increased ATX Activity Is Specific for Pruritus of Cholestasis.

Compared to healthy controls, ATX activity was slightly, but significantly, increased in patients with atopic dermatitis and Hodgkin's lymphoma (HL) and strongly increased in patients with cholestatic liver diseases (Fig. 1A). However, the strong elevation in ATX activity observed in patients with cholestasis with pruritus, compared to nonpruritic cholestatic controls, was not observed in age- and gender-matched cohorts of HL and uremia with versus without pruritus (Fig. 1A; Supporting Tables 1-3). Because all patients with atopic dermatitis suffer from itch, this comparison could not be made for this disease group. Strongly increased ATX activity appears therefore specific for pruritus of cholestasis.

Figure 1.

Elevated serum ATX activity is specific for pruritus of cholestasis. (A) Patients with atopic dermatitis, Hodgkin's disease, or cholestasis had significantly increased ATX activity, compared to healthy controls. In patients with Hodgkin's disease and uremia, there was no significant difference in serum ATX activity between patients with and without pruritus. In contrast, all patients with cholestasis with pruritus had significantly increased serum ATX activity, compared to patients with cholestasis without pruritus. *P < 0.05; ***P < 0.001 (ANOVA). (B) Increased ATX activity in patients with pruritus with cholestatic liver disorders was observed irrespective of the underlying disease. (C) ROC curve distinguishing patients with cholestatic pruritus from patients with pruritus resulting from atopic dermatitis, HL, or uremia. Sensitivity, specificity, and PPV were calculated using a cut-off value of 8.5 nmol·mL−1·min−1.

Our cohort of patients with chronic liver diseases suffering from pruritus consisted of PBC, PSC, BRIC, progressive familial intrahepatic cholestasis, chronic viral hepatitis C infection, CCC, hepatic sarcoidosis, liver cirrhosis, and drug- or toxin-induced intrahepatic cholestasis (Fig. 1B). Irrespective of the underlying cause of cholestasis, ATX activity was increased in all patients suffering from cholestatic pruritus. Enzymatic activity and itch intensity correlated linearly in this large group of patients (Supporting Fig. 1). In contrast, neither total bile salts nor FGF-19 levels showed any correlation with itch intensity in our patient cohort (data not shown).

Using a cut-off level of 8.5 nmol·mL−1·min−1 ATX activity had a sensitivity of 71%, a specificity of 91%, and a positive predictive value of 70% to diagnose pruritus resulting from liver disorders, in comparison to atopic dermatitis, uremia, or HL (Fig. 1C). Thus, in patients with pruritus of unknown origin (PUO) or in the case of a coexistence of two or more potentially pruritus-inducing disorders, ATX activity might be a useful diagnostic tool to identify patients suffering from a yet-undiagnosed liver disorder.

Colesevelam Marginally Lowers ATX Activity and Itch Intensity.

The current guidelines for the treatment of cholestatic pruritus recommend the use of bile salt sequestrants as first-line therapy.2, 4 In a recent double-blind, randomized, placebo-controlled multicenter study,12 however, colesevelam had only a mild effect in alleviating pruritus of cholestasis and was not more effective than placebo (Fig. 2A,B). As expected, bile salt levels were lowered in patients taking colesevelam (−49%; P < 0.01; Fig. 2A). This alteration was physiologically relevant, as shown by a similar reduction in circulating levels of FGF-19, the product of the bile salt receptor, farnesoid X receptor–stimulated FGF-19 gene (−47%; P < 0.01; Fig. 2A). ATX activity was slightly reduced (−13%) in the verum group (13.3 ± 5.6 nmol·mL−1·min−1 at baseline versus 11.6 ± 4.4 nmol·mL−1·min−1 after treatment; P < 0.05; Fig. 2B), whereas in the placebo group, ATX, total serum bile salts (TBS), and FGF-19 levels all remained unchanged (Fig. 2B).

Figure 2.

Marginal effects of colesevelam on ATX activity and itch intensity. Patients receiving either colesevelam (A) or placebo (B) were analyzed for itch intensity, ATX activity, TBS, and FGF-19 levels before and after 3 weeks of treatment. Colesevelam effectively reduced TBS and FGF-19 levels by approximately 50%, but pruritus was similarly reduced, compared to the placebo group.12 ATX activity slightly dropped in the colesevelam group.

RMP Attenuates Itch Intensity and ATX Activity.

When bile salt sequestrants are ineffective, RMP is recommended as second line therapy of cholestatic pruritus.2, 4 Six patients who did not experience improvement of pruritus using bile salt sequestrants were treated with 150 mg of RMP twice-daily. Itch intensity improved within 2 weeks of RMP treatment (−65%; P < 0.01; Fig. 3), which was accompanied by a concomitant significant decrease of ATX activity (−32%; P < 0.05; Fig. 3). TBS and FGF-19 levels remained unaltered during this treatment (Fig. 3).

Figure 3.

RMP therapy attenuated pruritus and reduced ATX activity. Patients with cholestatic pruritus not responding to colesevelam were treated with RMP for 2 weeks. Serum samples taken before and after treatment were analyzed for itch intensity, ATX activity, TBS, and FGF-19 levels. RMP attenuated itch severity and reduced ATX activity, whereas TBS and FGF-19 levels remained unaffected.

Rifampicin Does Not Affect ATX Activity, but Reduces ATX Expression in Human Hepatoma Cells In Vitro.

To elucidate the molecular mechanism of the antipruritic properties of RMP, we analyzed the effects of RMP on ATX activity and expression in vitro. RMP, at concentrations up to 100 μmol/L, did not modify ATX activity in serum (data not shown). Using HepG2 cells, however, RMP attenuated ATX gene expression in HepG2 cells (P < 0.01; Fig. 4A). Because RMP exerts its transcriptional effects through the PXR, we further analyzed its effect in HepG2 cells overexpressing PXR or after knockdown of PXR. In PXR-overexpressing cells, RMP caused a stronger inhibition of ATX transcription (P < 0.02; Fig. 4B), whereas this effect was lost in HepG2 cells after knockdown of PXR using shRNA (Fig. 4C). For all experiments, cytochrome P450 3A4 gene expression served as a positive control to verify the action of RMP (Supporting Fig. 4A-C). These data show that the expression of ATX is reduced at the transcriptional level by RMP and that this mechanism is mediated by PXR.

Figure 4.

RMP diminished ATX messenger RNA (mRNA) expression in vitro through a PXR-dependent mechanism. ATX mRNA expression was reduced (−25%; P < 0.01; n = 6) in HepG2 cells after incubation with 10 μmol/L of RMP for 24 hours (A). This inhibitory effect of RMP was increased (−40%; P < 0.02; n = 3) in HepG2 cells overexpressing PXR (B), whereas it was lost in PXR knock-down cells, but not cells being transduced with scrambled shRNA (C).

MARS Responders Show Diminished ATX Activity in Parallel With Reduced Itch Severity.

Severity of pruritus was evaluated in patients undergoing MARS therapy using VAS and a recently published itch severity score (ISS).16 The ISS showed a strong linear correlation with VAS (r = 0.92; P < 0.001; Supporting Fig. 2A). Eight patients had a marked improvement in itch intensity on VAS (−63.6%; P < 0.01) and ISS (−60.9%; P < 0.01; Supporting Fig. 2B) after MARS therapy and were designated “responders,” whereas 2 “non-responders” showed no change in severity of pruritus on VAS (−4.2%) or ISS (−2.2%) (Fig. 5A,B). A mean reduction of ATX activity of −29% (P < 0.01) was observed in responders, whereas nonresponders remained unchanged (Fig. 5A,B and Supporting Fig. 2C). The change in ATX activity directly correlated with the reduction in ISS (r = 0.71; P < 0.01; Supporting Fig. 2D) and VAS (r = 0.61; P < 0.03; Supporting Fig. 2E). TBS concentrations and FGF-19 levels (Fig. 5A) dropped in responders without reaching significance, whereas an apparent increase was observed in the 2 nonresponders (Fig. 5B). Neither ATX activity nor ATX protein was detectable in albumin dialysate (Fig. 5C,D), in line with the MARS membrane pores having a molecular-weight cutoff of 50 kD, which is approximately half the size of ATX. Intriguingly, ATX levels returned to pretreatment values with relapse of itching, which occurred in responders between 6 weeks and 4 months. Two patients underwent a second MARS treatment upon relapse of pruritus. During the second intervention, pruritus improved, again accompanied by a concomitant reduction of ATX activity (Fig. 5E).

Figure 5.

MARS therapy diminished itch intensity in most patients with refractory pruritus and caused a concomitant reduction in ATX activity. Patients undergoing MARS were divided into responders (A) and nonresponders (B) on the basis of VAS and itch severity score (Supporting Fig. 2). Pruritus and ATX dropped significantly in responders and remained unchanged in nonresponders. Neither ATX activity (C) nor ATX protein (D) could be detected in the albumin dialysate of MARS patients. (E) Individual courses of itch intensity and ATX activity in 2 patients undergoing sequential MARS therapy to treat otherwise intractable pruritus. Improvement of pruritus correlated with the reduction of ATX levels. Upon reoccurrence of pruritus, ATX levels had returned to pretreatment values.

Nasobiliary Drainage Strongly Reduced Pruritus and ATX Levels.

Nasobiliary drainage effectively alleviated intractable pruritus in PBC patients not responding to standard treatment.7 Simultaneously with the improvement of itch severity (−85%; Fig. 6A), ATX serum activity dropped in these patients to approximately half the baseline values (−50%; Fig. 6A), whereas TBS initially dropped, but rose back to baseline values already during nasobiliary drainage, as, in part, reported on previously7, 8 (Fig. 6A and Supporting Fig. 3A). Circulating FGF-19 levels were strongly diminished 1 day after the start of treatment, indicating effective external biliary drainage (−50%; Fig. 6A). Our observation that ATX activity closely correlated with improved itch intensity in patients undergoing nasobiliary drainage8 is strengthened by the reproducibility in 1 PBC patient who underwent this procedure twice (Fig. 6B). Because neither ATX protein nor ATX activity were detected in bile,8 the reduction in circulating ATX levels cannot be explained by the biliary clearance of ATX.

Figure 6.

Refractory pruritus and ATX activity were strongly diminished by nasobiliary drainage. (A) Treatment significantly reduced pruritus and ATX activity, whereas TBS were hardly changed. FGF-19 levels dropped significantly, indicating efficacy of the drainage procedure. (B) ATX activity and itch intensity in 1 patient undergoing successive treatment with nasobiliary drainage.

In summary, itch severity and ATX serum activity were barely reduced by colesevelam, moderately diminished by RMP and MARS therapy, and markedly diminished by nasobiliary drainage. The improvement of pruritus showed a linear correlation with the reduction in ATX serum activity for all treatment groups (Fig. 7A), whereas no correlation was found for the change in serum TBS concentrations (Fig. 7B).

Figure 7.

Overview of therapeutic interventions in patients with chronic cholestasis suffering from pruritus. Effect on itch intensity and ATX activity is shown for colesevelam, RMP, MARS, and nasobiliary drainage. Individual changes of itch intensity and ATX activity (A), but not TBS (B), after therapy (related to the pretreatment value) showed a linear correlation (Pearson's correlation coefficient: r = 0.62; P < 0.0001). Dashed line represents no change in itch intensity and ATX activity, respectively.


In the present study, we demonstrate that elevated serum ATX activity has a high specificity for pruritus of cholestasis and might therefore serve as a diagnostic marker in cases of PUO or multiple underlying diseases. A strong correlation between ATX activity and efficacy of pruritus treatment further strengthens the role of ATX in the pathogenesis of cholestatic pruritus. The beneficial effect of RMP on cholestatic pruritus may be explained, at least in part, by the PXR-dependent inhibition of ATX expression, as observed in vitro.

LPA is generated by ATX, and the serum levels of both correlate with the occurrence of cholestatic itch.8 Quantification of LPA can be artificially altered after blood sampling through release by platelets, and levels may vary depending on processing and storage.17 To circumvent these potential artifacts, we analyzed ATX activity as a reliable parameter for LPA formation. The source of the increased circulating ATX levels remains elusive, but might either be the result of reduced clearance, increased expression, or a combination of both. A reduced clearance may result from decreased uptake by liver sinusoidal endothelial cells.18 Despite their completely different mechanisms of action, RMP, MARS treatment, and nasobiliary drainage all markedly reduced ATX serum levels, whereas ATX protein was neither directly drained into bile8 nor removed in albumin dialysate. We hypothesize that a factor capable of increasing ATX expression (or reducing its clearance) is removed by these treatments. This yet-to-be-identified factor might accumulate in the circulation during cholestasis and might be metabolized in the liver and/or the gut, followed by biliary secretion and reabsorption through the enterohepatic circulation. The different therapeutic approaches might intervene at different stages in this cycle.

Colesevelam binds various amphiphilic substances in the gut lumen and was believed to effectively improve pruritus in patients with cholestasis. The binding capacity of colesevelam for the ATX-inducing factor might be minimal, as opposed to that for bile salts, which is underlined by only a small, though significant, decrease in ATX activity. Because cholestyramine has been reported on in uncontrolled trials to attenuate pruritus, it might be that cholestyramine could bind the ATX-inducing factor better than colesevelam, which was not superior to placebo in diminishing pruritus.12

RMP alleviates pruritus in cholestasis by, so far, unknown molecular mechanisms. Our in vitro data suggest that the antipruritic action of RMP in cholestasis can be explained, at least in part, by the transcriptional inhibition of ATX expression in a PXR-dependent fashion. This may explain why RMP is effective in pruritus of cholestasis, but not in pruritus of other origin, such as uremia, Hodgkin's disease, or atopic dermatitis, where systemic ATX does not play a major pathogenetic role. One could speculate that, in addition, RMP may reduce an ATX-inducing factor by the modulation of PXR-regulated genes involved in hepatic and intestinal detoxification and secretion and/or by alteration of the gut flora.4 Determination of plasma ATX activity and LPA levels in animal models of cholestasis in the presence and absence of an effective PXR agonist may teach us more about ATX and LPA turnover under these pathological conditions in the future. Alternatively, one has to consider that RMP might exert antipruritic effects, at least in part, by PXR-independent mechanisms. An experimental approach to test this option could be to compare the scratch response of mice toward injection of LPA with or without previous administration of rifampicin—a PXR agonist in men, but not in mice.

MARS therapy removes countless undefined substances from the circulation,5 possibly including the ATX-inducing factor. Nasobiliary drainage removes secreted bile from the body and thereby, possibly, also removes the ATX-inducing factor from the enterohepatic circulation. Further in vitro analyses in cell-culture systems of bile or albumin dialysates of patients with pruritus undergoing nasobiliary drainage or MARS treatment, respectively, could possibly help to identify the ATX-inducing factor in cholestatic pruritus.

It is of note that in as much as 10%-35% of patients presenting with chronic generalized pruritus, an internal disease can be determined as underlying cause.19 Despite extensive diagnostic examination, the cause of itching could not be identified in 8%-20% of patients with generalized pruritus.20-22 Eisendle et al. reported, in a study with 117 patients with PUO, that almost 30% of these patients had elevated TBS concentrations without any evidence for liver disease.22 Identifying the underlying disease causing pruritus apparently is a clinical challenge, and diagnostic parameters are warranted to make a differential diagnosis. ATX may represent such a novel marker for pruritus of cholestasis. In this study, an increased enzymatic activity above 8.5 nmol·mL−1·min−1 had a positive predictive value (PPV) of 70% in differentiating cholestatic pruritus from pruritus associated with atopic dermatitis, uremia, and HL. Determination of ATX serum activity in PUO or, more important, in cases of the coexistence of two or more potentially pruritus-inducing disorders might help clinicians in choosing a targeted therapeutic regimen.

Slightly increased serum ATX activities were observed in patients with atopic dermatitis and HL, compared to healthy controls, in our cohort. A local overproduction of ATX with only marginal increases in the systemic circulation could be a conceivable mechanism causing itch perception in these patients. In line with our results, slightly enhanced ATX levels have been reported in a small cohort of 11 HL patients, compared to healthy controls.23 In relation to uremia, it has been reported that patients with renal failure have 3-fold elevated circulating LPA levels, compared to control subjects.24 More recently, this was confirmed in a rat model of unilateral urethral obstruction.25 Strikingly, in this study, elevated plasma LPA was accompanied by increased ATX activity in renal effluent, rather than in plasma. It could be hypothesized that in renal failure, ATX is primarily secreted into primary urine, but its product, LPA, may also end up in plasma. Hence, this leaves open the possibility that LPA plays a role in itch perception also in atopic dermatitis, Hodgkin's disease, and uremia.

Several experimental and clinical observations favored increased levels of bile salts as causative pruritogens in hepatobiliary disorders in the past.26 However, no correlation between the level of any naturally occurring bile salt in the circulation or skin and severity of pruritus could be proven.26 In addition, several observations in the present study argue further against a direct causal role of bile salts in pruritus: (1) Colesevelam halved TBS levels without being more effective than placebo regarding improvement of itch intensity; (2) RMP or MARS therapy did not significantly reduce bile salt levels, yet strongly diminished itch severity; (3) in patients undergoing nasobiliary drainage, TBS levels dropped initially, but returned to baseline values during treatment long before pruritus reoccurred; and (4) the lack of correlation between TBS concentrations and itch perception.

In our patient cohorts, markedly elevated ATX activity was specific for pruritus of cholestasis. Thus, ATX might represent a useful diagnostic tool for those cases in whom chronic pruritus remains unclassified. In addition, our study provides further clinical and experimental evidence that ATX inhibitors and LPA-receptor blockers may have potential as future therapeutic agents to effectively treat pruritus in cholestatic liver disorders.


The authors thank Dagmar Tolenaars for her expert technical assistance.